Convolutional Neural Network based Matchmaking for Service Oriented System Construction

Abstract

In recent years, service-oriented computing, as a new computing paradigm, has been developed rapidly. Following this trend, more and more Web services and cloud services have been developed and made publicly available on the Web. These publicly available services will be important components for service-oriented system construction. However, these large numbers of services increase the burden of service selection in service-oriented system construction. This paper proposes a convolutional neural network-based service matchmaking approach to match services according to developer requests in service-oriented system construction. In the model, the convolutional neural network aims to learn the semantic feature representation for matchmaking. We experimented on a real-world dataset, ProgammableWeb.com, and experiment results show that the proposed approach can help find relevant services according to developer requests.